It is the central hypothesis of this proposal that adenovirus (Ad)- mediated transfer of genes coding for tumor antigens to autologous DC, is an effective strategy to evoke tumor-specific CTL, and consequently suppress the growth of tumors. Further, we hypothesize by modifying DC or their milieu by Ad-mediated expression of specific cytokines associated with their development and function, DC antigen presentation and T-cell stimulation can be augmented. We base these hypotheses on the unique properties of Ad vectors, the knowledge that genetically modified antigen -presenting cells favor CTL induction since the transgene product is present through the MHC class I system, and our preliminary data supporting the basic concepts. To evaluate these hypotheses, this study has four specific aims: (1) to evaluate the ability of 1st generation Ad vectors and vectors genetically modified to enter cells through alternative receptors, to transfer and express genes in murine DC ex vivo, and the trafficking and survival of the genetically modified DC following in vivo administration to syngeneic mice; (2) using Ad vectors to transfer genes coding for tumor antigens to murine DC ex vivo, to evaluate the consequences of single and repetitive in vivo administration of the genetically modified cells to generate CTL directed against the tumor antigens and the suppression of growth of tumors expressing these antigens in syngeneic mice; (3) using the principles derived from the murine system to optimize the generation of human CD34+- derived DC Ad-mediated transfer of transgenes to human CD34+-derived DC, and to evaluate the ability of human DC modified with genes coding tumor antigens to stimulate autologous T-cells to evoke the generation of tumor antigen-specific CTL; and (4) based on the data generated in specific aims 1-3, to carry out a human trial of Ad-mediated transfer of genes coding for tumor specific antigens to individuals with metastatic melanoma, the study will compare the ex vivo DC strategy to direct in vivo gene transfer.